One aspect to operational risk, which can have a meaningful impact on performance, not to mention the integrity of the manager, is valuations. According to Deutsche Bank's Global Prime Finance group's third annual Operational Due Diligence Survey, asset valuation is in sharp focus for 38 per cent of investors surveyed. Unilaterally, respondents said they would review a fund's valuation policy during their ODD review whilst 78 per cent stated they would verify the valuation procedure during the on-site review.
Asset valuation is inherently a function of the quality of data inputs. A SunGard survey last year revealed just how important risk pricing is in the front office. Some 20 per cent of respondents attributed its importance to calculating credit valuation agreements (CVAs), 31.8 per cent to initial margin requirements and 22.7 per cent to funding valuation adjustments and other funding costs. Admittedly, this survey centred on banking entities but the issues apply equally to the asset management industry.
"Operational risk management is something I always think about holistically; market, liquidity, credit, operating and residual risks," says Andrew Kandiew, Head of Operational Due Diligence at K2 Advisors, a leading fund of hedge funds manager with USD10.3 billion in AUM. "This is risk management but one that focuses more on qualitative aspects when evaluating the manager.
"Operational due diligence has become much more important post-08. Your ODD has got to have teeth and you've got to demonstrate that it works. In terms of pricing and valuation, we will always seek to fully understand the investment strategy of any manager; what they are trading and the instruments they hold in their portfolio. From there, we consider a manager's pricing and valuation risk. We look at a host of operating risk factors, and valuation is a key one.
"For example, we will look at whether a manager has written valuation procedures that are germane to the strategy, as well as a governance structure around the valuation process i.e. does the manager have a valuation committee? What is its composition? How are issues escalated? Who are the people and infrastructure used to support pricing and valuation? Are there clear segregation of duties within the firm?" explains Kandiew.
Such is the importance of operational risk that strategically, K2 Advisors uses a completely independent approach whereby Kandiew does not report to the head of research. Research is charged with the responsibility of looking for potential investment managers and whilst there is a spirit of collaboration within K2 Advisors, "my role is independent of the research team. I have a veto that I can exercise, and at times do, when they recommend managers to invest with," adds Kandiew.
Model risk
The accuracy of a manager's valuations ultimately depends on the quality of data inputs but it also depends on the valuation model being used. Valuation sounds like it should be precise but valuation models rely on using different assumptions such that it can only ever be an approximation. There is some acknowledgement that mark-to-market and mark-to-model techniques do not necessarily provide accurate numbers.
The level of uncertainty that comes with using a valuation model is known as model risk and it is something that risk managers have become far more focused on recently. Knowing the limitations of a model has been an important step change.
A big part of operational risk could therefore stem from valuations. To make this more strategic, risk managers should review their models and the implicit assumptions that underpin those models on a periodic basis; stress testing, sensitivity analysis etc.
"We service a number of private debt strategies and when it comes to doing valuations you have to do a great deal of sophisticated credit analysis (model behaviours, weighting assignments etc)," says Alan Picone, global head of risk consulting and infrastructure at Duff & Phelps' Kinetic Partners division.
Model risk is something that Dr Lawrence Wormald, COO and head of research, SunGard APT, gets asked about a lot.
Take the Black-Scholes model for pricing options as an example. One of the greatest financial innovations ever, the assumptions used in Black-Scholes are actually very simple and as such bear no resemblance to reality. This has led to Black-Scholes being applied back to front.
Instead of saying, "I think this is the volatility and therefore this is the price of the option", SunGard APT uses the market to price the option and then works backwards to determine what the volatilty should be. This is referred to as implied volatility.
"We have to make sure that we can justify the implied alpha when clients look at their portfolio. We believe in Black-Scholes, precisely because we use it backwards not forwards; which would otherwise create model risk," explains Wormald.
One important development to help improve risk management is to integrate liquidity modelling into risk models. This is something that SunGard has recently rolled out within APT.
When there are shocks to prices there are often shocks to liquidity as well and it becomes harder to close out positions, or indeed put on positions. That ability to combine price and liquidity shocks in the model has therefore become critical.
"The key is to provide intiution rather than just an arbitrary number. A few years ago people came up with liquidity-adjusted VaR, which was always a bit bigger than the ordinary VaR. But why? Few could understand it. We are trying to improve on that by using a factor model for liquidity risk. It is already being consumed by five or six of our global clients in APT. It satisfies the regulator but more importantly it gives managers extra information to make more informed investment management decisions," confirms Wormald.
Protect yourself from failure of imagination
At SkyBridge Capital, managing risk has always been strategic. It has put significant effort into integrating risk management into portfolio construction to ensure that time and again, the investment portfolio minimises losses on the downside and maximises gains on the upside. Applying a disciplined approach has allowed the SkyBridge flagship fund to outperform in every market sell-off over recent times.
"There are a lot of levers that go in to managing risk. We take a holistic approach to risk management, and continually enhance and broaden our analytical capabilities to aid our portfolio management in generating desired risk-adjusted returns. We are always adding new tools to our risk management tool kit," explains SkyBridge Partner and Head of Risk Management, Tatiana Segal.
SkyBridge's approach to managing risk is to not only rely on looking in the rear-view mirror but to be as forward-looking as possible. This is achieved by using Imagine Software to provide position-level detail within the portfolio in tandem with proprietary tools.
"Broadly speaking, we focus on embedded alpha, while our beta will fluctuate depending on our conviction, with historical ranges around -0.2 to +0.5. Our approach is to continually analyse multiple scenarios of how the risk/reward can play out so that we stay best positioned to extract our targeted compensation, over a full market cycle, for each unit of risk that we take. The Beta expression will change accordingly," says Segal.
"We test the portfolio using both historical and hypothetical scenarios, and while we continually add new scenarios relevant to the current market environment, we remain mindful of what a truly worst case might be. Although we remain constructive, it's important to not grow complacent. That is why we focus on modeling portfolio liquidity assuming multiples of 2008 redemption levels, the highest-ever redemptions we have seen since inception.
"I've heard it said that one of the main problems with risk management frameworks pre-crisis was a failure of imagination. We are trying to learn from that. We built our risk management platform to be rooted in the empirical data, yet forward looking that consider a wide spectrum of possibilities. It's a delicate balance."
One of the ways that Duff & Phelps' Kinetic Partners division is helping managers to improve their risk management activities, particularly in response to market regulations and reporting obligations under the AIFMD, is to provide an advisory service as part of a wider AIFM solution.
This involves performing a number of model validations – be it algorithms that are used by systematic CTAs, be it risk systems or valuation models.
"We have a proprietary solution for risk reporting; it is a full fledged risk management and risk measurement solution. Typically, hedge fund managers can send us their portfolio which we process in our system and we produce for them the Annex IV report. It captures everything from sensitivities to conditional VaR, expected shortfall and so on.
"This is ideal for managers who want the report first thing in the morning to analyse. Typically this is being done on a daily basis," confirms Picone.
Any time a hedge fund manager uses risk systems, there is often a disconnect between the front office traders and the middle office team, which might be generating a daily report on risk that the investment decision makers do not agree with.
To make risk more strategic, therefore, requires a system that provides intuition as well as an enterprise-level capability. By doing so, managers are able to tell a more coherent story about how they manage risk, thereby enhancing their reputation and asset raising capabilities.
"We think we can help our clients achieve that because we give them a comprehensive and coherent model in APT. Coherent in the sense that a single risk model is designed to provide coverage across all major asset classes. It's also intuitive because APT is a factor model. It's not just based on time series and pure statistical approaches," explains Wormald.
Whatever the story is, you need to have a system that people internally buy-in to as well as externally when speaking to prospective investors. That's when risk can become a part of the strategic vision of an asset management business.
Having a risk system that makes traders feel that they are getting good information is important.
"Investors gain intuition by thinking about risk in factor terms. Most investment professionals believe that there are different factors driving the market; it might be interest rates, commodities pricing, a restructuring story within a market segment. If you can relate risk to those factors that are driving the market, then people believe that the risk numbers you use make sense. They might then have the confidence to take on more exposure because they have a better understanding, and potentially boost returns.
"If you don't have a factor model you won't be able to get insights from asking the system, `Tell me about my China exposure', or `Tell me about my oil exposure'. For a straightforward European long/short equity strategy, the implied exposure you have to commodities or export companies that do business in China will be a risk factor and a potential return driver. If you're not capturing that with your risk system, it's impossible to be strategic," stresses Wormald.
The wisdom of the crowd
The Alternative Share Private Network (`ASPN Solutions') is a comprehensive front, middle, and back office solution designed to empower institutional investors to build, execute and monitor multi-manager hedge fund portfolios. Dr Ken Akoundi, one of the founders of RiskMetrics and who joined ASPN last year, explains that the system has been built on the basis of the `wisdom of the crowd'.
As such, ASPN is a solution designed to help institutional investors collaborate with each other, share ideas, and basically get a clearer handle on the risk and performance attribution of managers in their portfolio. The reason why this is enabling investors to become more strategic is because ASPN provides them with a central repository to collect and store data. In Akoundi's view, institutions are less interested in granular position-level data. The hurdle many are still trying to overcome is getting all of their manager data in one place.
"They might have a document sharing system, a performance measurement system, a factor model system; that's three systems already. With ASPN they have one system with everything in one place. They've got their data in good order on the long only side but when it comes to hedge funds it's still a complete mess for most institutional investors.
"For many investors, extracting information from monthly newsletters, summaries and putting it all into a spreadsheet is timely. What we can do is create a data map for every manager letter. The day after a manager releases their monthly letter, it's in the ASPN system ready for the investor to review," says Akoundi.
When asked whether the collaboration capabilities of ASPN could change the way that institutional investors approach risk management, Akoundi responds: "Overall, I don't think using ASPN will change investors' risk appetite. They just want the ability to use risk tools that will tell them what will happen if something goes wrong. In my opinion, stress testing and sensitivity are two important concepts that are becoming more important than VaR. It's not only knowing how a 5 standard deviation left tail event would impact the portfolio, but also a 4 and 3 standard deviation event, a median standard deviation, and perhaps a 5 standard deviation right tail event. Users can do that visually in ASPN.
